论文部分内容阅读
在球不变随机向量的非高斯背景下,针对估计协方差矩阵可能奇异的情况,研究了距离扩展目标的自适应检测方法。首先,推导了非高斯背景下未知协方差矩阵和目标散射点幅度的修正最大似然(maximum likelihood,ML)估计;然后,基于纹理分量的近似ML估计,建立了自适应检测器(adaptively modified generalized likelihood ratio test,AMGLRT)。仿真结果表明,AMGLRT在目标散射点能量均匀分布时检测性能最佳,随着杂波尖峰的减小或阵元数的增加,AMGLRT的检测性能有所改善;且其对不同杂波相关性表现出很好的鲁棒性。另外,AMGLRT的检测性能优于已有的M/K检测器,且这种性能优势随着散射点个数的增加而增大。
In the non-Gaussian background of the ball-invariant random vector, aiming at the possible singularity of the estimated covariance matrix, the adaptive detection method for distance-expanded targets is studied. Firstly, the maximum likelihood (ML) estimation of the unknown covariance matrix and the target scatterer amplitude in non-Gaussian background is derived. Then, based on approximate ML estimation of texture components, an adaptively modified generalized likelihood ratio test, AMGLRT). The simulation results show that AMGLRT has the best detection performance when the energy of target scattering point is uniform and the detection performance of AMGLRT improves with the decrease of clutter spikes or the number of array elements. Out of good robustness. In addition, the detection performance of AMGLRT is superior to the existing M / K detector, and this performance advantage increases as the number of scattering points increases.